Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.

[next] An Internet archive server server (was about Lisp)

Skip to first unread message

Peter Deutsch

Sep 11, 1990, 1:45:06 PM9/11/90
Archive-name: archie/10-Sep-90
Original-posting-by: (Peter Deutsch)
Original-subject: An Internet archive server server (was about Lisp)
Archive-site: []
Reposted-by: (Edward Vielmetti)

In article <49...@brunix.UUCP>, (Axel Merk) writes:
> help! I NEED LISP!!!

[ And several other people wanted a Lisp ]

This posting is a follow-up to all those poor souls who
wrote about not having a Lisp for their machine after the
"big announcement" and how they would immediately sell
their NeXT and go ice fishing if they are not given one.

Well, I'm still steaming about the reaction to an earlier
posting, so let me make it ABSOLUTELY CLEAR. I DON'T

Fine, with that out of the way, I am writing this to tell
you about:

a) a (possibly useful) new tool for net users.
b) A nice version of Lisp, and where to get it.

First, the nifty new tool for net users. For some time we
have been faced with a problem many of you will be familar
with. We knew there is a lot of good free software out
there, but couldn't always find what we wanted. We started
to build up a list of archive sites, but logging in around
the net looking for things didn't seem to be a productive
way to spend our time. The problem was not having or finding
working archive servers, it was keeping a current list of
what was on them.

Thus was born "archie", an Internet archive server server
(or a server^2 :-)

Credit for this idea is sharedby me, Alan Emtage
(, who set up the ftp search software,
and Bill Heelan ( who wrote the
interface program currently in use.

Alan has a set of shell scripts that automagically calls
up some subset of our list of servers each night (to keep
load down on any one machine) and do a remote 'ls' on that
machine. We cycle through the entire list in about a
month, so no entry can be more than that old. For what
it's worth, we currently know of about 210 sites.

This was originally used only by my staff, grep'ing
through the listings for strings. We currently store
everything in a series of files (eew!), one per host. We
then NFS'ed the appropriate directory to our student
machines and made it available to them. Finally, after a
number of requests from people on the net who had gotten
wind of this, we have prepared a "pre-beta" version of

Archie allows anyone to rlogin into one of our machines
and query our archive lists. Another one of my guys, (Bill
Heelan, wrote a bare-bones front
end, which allows people to login, get a help message, ask
for a specific string (using ed-like regular expressions)
or get the contents of a specific archive server.

Currently, archie is pretty brain-damaged. The files are
stored as flat ASCII files and the "prog" command just
launches a "grep" with the appropriate arguments. Future
plans (once we get the latest start of semester out of the
way) is to add a real database (we've compiled postgres
(sp?) and Alan has started to code the program to parse
the raw input). This will dramatically increase
performance (the current system greps through about 31
Meg of listings!!). We are also looking at a number of
other enhancements, including a mail-based interface so
UUCP sites can write to us with their queries. Still, the
basic functionality is there now.

So, if you are looking for some public domain software, be
it kcl or DES encryption, you are welcome to give archie a
try. The current version runs on a Sun 4/280, which will
see a lot of use this semester, so we might have to put
certain restrictions on it later. For now, we are
soliciting input, both on the original idea and our
brain-damaged implementation.

Oh, yeah. How to reach us. If you have Internet access,
try an "rlogin -l listd" (I hope to
get this transfered to "archie" instead of "listd" any
day, but the old name will live for compatibility for a
while). You should get a message with instructions and
then the prompt. You can then ask for a specific site, or
run the grep (command is "prog <pattern>").

Please don't go overboard. If you want our list of archive
sites, or a copy of the software that lets us do this,
don't run a 'prog *", write to either Alan, or Bill. We
will probably take a while to answer, so bear with us.
Meanwhile, your feedback on the idea is welcome.

BTW, strictly speaking this probably is not a NeXT issue,
as we are not currently running it on a NeXT (that could
change soon, more details later). Meanwhile, it might help
calm those of you about to sell your NeXT machines.
There's a lot of free software out there. I hope this
helps you find some of it.


Now, the public domain version of Lisp I mentioned in
the posting before this one. I have already received mail
asking for the name of an archive site, so here is a
minimal subset of archive servers that claim to have kcl
(Kyoto Common Lisp). This was gleaned by letting archie
run for a few minutes, stopping when I had enough.

These are all ftp-able archive sites which claim to have
some version of kcl available. Some have a single huge tar
file, some have it nicely broken up into parts. I remember has it in pieces, for example. One should be able
to help you.

Don't forget to send in your license form. Enjoy.

- peterd

+-------+ Peter Deutsch McGill University
| u # u | School of Computer Science
| a a |
\ a / "Love my work, hate my job..."

0 new messages